The bivalent ligand, MMG22, reduces neuropathic pain after nerve injury without the side effects of traditional opioids

Abstract

Functional interactions between the mu opioid receptor (MOR) and the metabotropic glutamate receptor 5 (mGluR₅) in pain and analgesia have been well established. MMG22 is a bivalent ligand containing MOR agonist (oxymorphamine) and mGluR₅ antagonist (MPEP) pharmacophores tethered by a 22-atom linker. MMG22 has been shown to produce potent analgesia in several models of chronic inflammatory and neuropathic pain. This study assessed the efficacy of systemic administration of MMG22 at reducing pain behavior in the spared nerve injury (SNI) model of neuropathic pain in mice, as well as its side effect profile and abuse potential. MMG22 reduced mechanical hyperalgesia and spontaneous ongoing pain after SNI, with greater potency early (10 days) as compared to late (30 days) after injury. Systemic administration of MMG22 did not induce place preference in naïve animals, suggesting absence of abuse liability when compared to traditional opioids. MMG22 also lacked the central locomotor, respiratory, and anxiolytic side effects of its monomeric pharmacophores. Evaluation of mRNA expression showed the transcripts for both receptors were co-localized in cells in the dorsal horn of the lumbar spinal cord and dorsal root ganglia. Thus, MMG22 reduces hyperalgesia after injury in the SNI model of neuropathic pain without the typical centrally mediated side effects associated with traditional opioids.

Summary

MMG22 potently reduced mechanical hyperalgesia and spontaneous pain early after nerve injury, while lacking rewarding properties and central side effects of traditional opioids.

1. Introduction

Chronic pain affects an estimated 100 million American adults, at a cost of approximately $600 billion USD annually [55]. Despite this economic and social burden, few new treatments have become available to combat this problem. Neuropathic pain (NP) is pain caused by a lesion or disease affecting the somatosensory system [57]. NP can be caused by traumatic injury and is also associated with a number of common conditions including: diabetic neuropathy, post-herpetic neuralgia, chemotherapy-induced peripheral neuropathy, neuropathic cancer pain, HIV associated neuropathy, and trigeminal neuralgia [121]. The prevalence of NP is thought to be ~7–10% of the general population [17,31,49,142]. NP is characterized by allodynia, hyperalgesia, paresthesias, and spontaneous ongoing pain [27].

Although non-steroidal anti-inflammatory drugs (NSAIDS) and opioids remain the best options and the mostly commonly prescribed medications for chronic pain [54], NP patients typically do not respond well to these analgesics [80,88,162] and less than 50% of patients receive adequate pain relief from currently available treatments [36,60]. Inadequate treatment for NP and the current epidemic of opioid addiction and overdose-related deaths underscore the need for new treatment strategies.

Earlier studies showed that co-administration of morphine and a metabotropic glutamate receptor 5 (mGluR₅) antagonist enhanced the antinociceptive effects of morphine and reduced its analgesic tolerance and dependence [69,91,131,166]. Functional interaction between mu opioid receptors (MOR) and mGluR₅, and evidence that MOR/mGluR₅ can form heteromers [122], led to the development of MMG22 [2]. MMG22 is a bivalent ligand that combines a mu opioid receptor (MOR) agonist (oxymorphamine) with a metabotropic glutamate receptor 5 (mGluR₅) antagonist (2-methyl-6-(phenylethynyl)-pyridine (MPEP) [42].

The two pharmacophores are tethered together with a 22-atom linker. Intrathecal (i.t.), but not intracerebroventricular, administration of MMG22 was orders of magnitude more potent at reducing lipopolysaccharide (LPS) or bone cancer-induced hyperalgesia than bivalents with shorter or longer linker lengths or a mixture of the monovalents [2,130]. Inflammation also causes a dramatic left shift in the analgesic dose response curve of MMG22 [2]. The analgesic potency of MMG22 may also depend on the ability of mGluR₅ to modulate N-methyl-D-aspartate (NMDA) receptor activity [3], which is integral to the development of synaptic plasticity and central sensitization [71,108].

The importance of linker length to the potency of MMG22 may depend on the pain model employed, the time after nerve injury at which it is given, its underlying mechanism, and the route of administration. For example, i.t. administration of the MMG bivalents with 10 or 22 atom linker lengths were equipotent in the spared nerve injury (SNI) model [107], whereas the i.t. potency of MMG22 was a million-fold greater than that of MMG10 in alleviating pain in mice with cisplatin-induced neuropathic pain (in preparation). Importantly, repeated administration of MMG22 did not promote the development of analgesic tolerance [127,130].

In the present study, we examined the antihyperalgesic effects of systemic administration of MMG22 delivered early or late after nerve injury and determined if MMG22 produced side effects typically associated with opioids or mGluR₅ antagonists such as abuse potential, hyper-locomotion, respiratory depression, constipation, and anxiolysis.

2. Methods

2.1. Animals

Adult (5–8 months) male and female C57/B6 mice (Charles River) were housed 4 (males) or 5 (females) to a cage and maintained on a 12-hour light/dark cycle with ad libitum access to food and water, except as otherwise noted for constipation studies. All procedures were carried out during the light cycle. 8–12 mice (an equal number of male and female mice) were used for each experiment unless otherwise specified. All procedures were approved by the Institutional Animal Care and Use Committee of the University of Minnesota.

2.2. Spared nerve injury

Mice were anesthetized with 2.5 % isoflurane. Spared nerve injury (SNI) to the sciatic nerve was performed as described previously [18,29,107]. Briefly, after exposing the three branches of the sciatic nerve, the tibial and common peroneal branches were tightly ligated with 5.0 silk suture and cut 2 mm distal to the suture. Care was taken not to disturb the sural nerve. Sham surgeries followed the same procedure without manipulation of the sciatic nerve or distal branches.

2.3. Drugs

The bivalent ligand MMG22 was synthesized as described previously [2]. MMG22, morphine, 2-methyl-6-(phenylethynyl)pyridine (MPEP), (Mallinckrodt Inc, Hazelwood, MO) and loperamide (Sigma, St. Louis, MO) were diluted in 1% DMSO (vehicle). Loperamide was diluted in near boiling 1% DMSO daily. All drugs were administered subcutaneously in a volume of 250 μl, between the shoulders. For all experiments, the experimenters were blinded to drug by a third party.

2.4. Behavioral measures of hyperalgesia

Mechanical hyperalgesia was defined as an increase in frequency of paw withdrawal evoked by a von Frey monofilament as we described previously [62,63]. Mice were placed on an elevated mesh platform under glass enclosures and allowed to habituate for 30 minutes prior to initial testing. A calibrated von Frey monofilament with a bending force of 5.9 mN (0.6g) (StoeltingCo, Woodale, IL) was applied to the lateral portion of the plantar surface of each hind paw 10 times, with an interval of approximately 10 seconds between applications, and the frequency of withdrawal responses were determined. Mice were tested a minimum of three times prior to and after surgery before analgesic testing. Drug-induced reduction of mechanical hyperalgesia was assessed early (10 days) and late (30 days) after surgery. 10 days was chosen as the early time point because full development of hyperalgesia occurs around 7 days after nerve injury [18,29] and 3 days of stable post-surgical baselines were desired prior to analgesic testing. We chose 30 days for the late time point because at this time, the inflammatory response to nerve injury has subsided [5,88,149)

2.5. Dose-response functions for antihyperalgesia

To determine dose-response functions, mice were injected with escalating doses of MMG22, morphine, loperamide or MPEP (s.c.). Separate groups of mice were used for each drug and time point. Starting dose of MMG22, morphine and loperamide was 0.1mg/kg, starting dose of MPEP was 1mg/kg. Doses were increased as follows: 0.1mg/kg, 0.3mg/kg, 1mg/kg, 3mg/kg and so on. The frequency of withdrawal evoked by 10 applications of a von Frey monofilament was determined 30 minutes after each injection. Mice were returned to their home cage after each injection and placed back on the mesh platform 10 minutes prior to testing. The percent maximal possible effect (%MPE) was calculated using the following standard formula:

$$\%MPE=\frac{\left(PostDrugValue-PreDrugValue\right)}{\left(PreSurgicalBaselineValue-PreDrugValue\right)}\times 100$$

Only doses that resulted in a behavioral response >0% were included in the analysis, The final dose included in the analysis was either the first dose to give a 100% MPE or the largest dose tested. The dose that reduced the withdrawal frequency by 50% compared to baseline (ED₅₀) was determined by non-linear regression of the %MPE data carried out in Prism 8.00 (GraphPad Software, San Diego, CA).

2.6. Conditioned place preference

The conditioned place preference (CPP) test was used to evaluate the rewarding properties of drugs in naïve animals [10,145,146]. A variation of this assay, the analgesic conditioned place preference (aCPP) test, was used to assess the effects of MMG22 on spontaneous, ongoing pain [28,65,137]. The CPP apparatus consisted of a two chambered box (28 × 28 × 20 cm) made from Plexiglas lined with 16 infrared photobeam emitters and detectors (Med Associates, St. Albans, VT). Alternate sides of the box were lined with vertical or horizontal black and white stripes (1.2 cm thick) but were otherwise identical. The two sides of the box were separated by one of two plexiglass partitions; a closed partition, or an open partition with a centrally located opening to allow access to both chambers. Both partitions had the same black and white striped patterns. On day 1, mice were placed into the middle of the box (with the open partition) and allowed to move freely between the two chambers for 30 minutes. Movement was tracked by software (Med Associates, activity monitor) that recorded infrared beam interruptions to locate the mouse position in the box. The program recorded the time spent in each chamber on days 1 and 5 of testing. Day 1 established if mice exhibited a preferred side of the chamber at baseline. Mice that spent more than 70% of the 30 minutes in one chamber were excluded to avoid preconditioning bias (1 mouse). On days 2–4 the remaining mice were subjected to two separate 30 minute chamber/treatment pairings per day. In the morning, mice were given a s.c. injection of vehicle and then placed in one side of the chamber 30 minutes later. In the afternoon, the same mice were given a s.c. injection of drug (at the same volume) and placed in the alternate (drug-paired) side of the chamber 30 minutes later. The drug paired chamber was pseudo-randomly assigned to each mouse such that some mice received drug in the chamber with vertical stripes, and some in the chamber with horizontal stripes, while maintaining a baseline average of ~40% of time spent in -in the drug paired chamber (biased design). During these sessions, the closed partition was used to separate the two chambers so that mice only had access to one side of the box. On day 5 (post-conditioning) the open partition replaced the closed partition and the mice were allowed to move freely between to two chambers for 30 minutes. The time spent on each side was again recorded. Preference scores were generated by subtracting the amount of time (sec) mice spent in drug paired chamber during pre-conditioning from the time mice spent in the drug paired chamber post-conditioning. Each mouse was only used for one conditioning experiment.

2.7. Sedation and anxiety

To determine if treatment with MMG22 shared any of the motor side effects common to centrally acting opioids we measured the effects of MMG22 on motor behavior. Mice were placed in an activity chamber box 30 minutes after drug administration and motor activity was recorded for the subsequent 30 minutes. Total distance traveled, average velocity, and total ambulatory time were recorded and analyzed.

Antagonists to mGluR₅ are known to decrease basal measures of anxiety in mice [56,135,148]. To determine if MMG22 shared any of these effects we used an open-field assay to look for center avoidance. Center avoidance is considered anxiety-like behavior in rodents [114]. Pharmacological studies have supported this interpretation as anxiolytics have been shown to increase both the total time mice spend in the center area as well as the number of entries mice will make into the center area during a given time [114,128,129,144]. Data sets were analyzed by defining a central area to be one-half the size of the full chamber, with identical central coordinates. Total time spent in the center, time spent ambulatory in the center, distance traveled in the center, and the number of entries into the center zone were recorded and analyzed.

2.8. Drug-induced constipation

The colonic bead expulsion test [118] was used to compare the effects of chronic MMG22 and morphine on constipation, a common side effect of opioids. Mice were given twice daily s.c. injections of vehicle or drug (MMG22 or morphine at 1 or 10 mg/kg) for 8 days. At 24 hours before testing, mice were placed in cages with raised mesh wire to suspend them above their bedding and prevent ingestion of feces or bedding. Mice were then fasted for 24 hours with free access to water; to maintain caloric intake and to avoid hypoglycemia, mice had access to a sugar water solution of 5% dextrose for the first 8 h of the fasting period. 30 minutes prior to testing, mice were given a final s.c. injection of vehicle or drug (MMG22 or morphine at 1 or 10 mg/kg). 30 min after injection, mice were anesthetized with 2.5% isoflurane (1–2 min) and a single 2 mm glass bead was inserted 3 cm into the distal colon. Bead insertion was accomplished by pushing the bead into the end of a 10 cm long piece of 2 mm plastic tubing. The tubing was inserted into the rectum 3 cm and then an internal plunger was advanced just past the tip of the tube to ensure bead ejection. After bead insertion, mice were placed in large glass beakers and the time to bead expulsion was recorded. Mice were monitored for a maximum of 5 hours.

2.9. Respiratory depression

Whole body plethysmography experimental setup was adapted from Young et al. (2018) [163] to determine the effects of morphine (3, 10 and 30 mg/kg) and MMG22 (3, 10, 30 mg/kg) on respiratory function. Mice were allowed to acclimate for two days in the same room as the plethysmography before testing. Mice were habituated to the testing environment by placing them in the chambers for 30 minutes for two consecutive days before testing. The mice were observed by the experimenter outside the room via camera for any adverse effects. On testing day, baseline respiratory data was collected for 30 minutes. Mice were given drug or vehicle and placed back in the chamber. Measurements of respiratory function were taken every 10 minutes for 60 minutes. Each mouse was utilized for three measurements on three separate days: Day 1: mice were injected with vehicle, Day 2: mice were given the lowest dose of either MMG22 or morphine, Day 3: mice received a higher dose of either MMG22 or morphine.

2.10. Localization of MOR and mGluR₅

RNAscope® in situ hybridization (ISH) (a probe based non-radioisotopic RNA ISH approach for detecting target RNAs in tissue) was used to determine whether MOR and mGluR₅ transcripts co-localized in dorsal root ganglia (DRG) or spinal neurons. Ten days after SNI surgery, mice were given an intraperitoneal injection of Euthasol (sodium pentobarbital, 390 mg/mL and phenytoin sodium, 90 mg/mL) and transcardially perfused with saline followed by 4% paraformaldehyde. L3-L5 DRGs and L3-L5 spinal cord segments were collected, post fixed in 4% paraformaldehyde for 3 hours and then placed in 30% sucrose in phosphate-buffered saline overnight at 4°C. Isolated and fixed DRGs and spinal cord segments were embedded into a tissue microarray using OTC media, frozen in dry ice and methanol, sectioned (7–10um thickness) with a cryostat, and thaw mounted onto slides. Sections were stored at −80°C.

RNAscope® ISH was performed on fixed, frozen sections of DRGs and lumbar spinal cord with probes for mouse Oprm1 (Cat No. 315848) and Grm5 (Cat No. 423631) purchased from ACD Bio. RNAscope®. ISH for sections was performed following manufacturer’s written protocol with only one modification to the protease digestion (ACD, RNAscope® Multiplex Fluorescent Detection Reagents v2, Cat No. 323110). Tissue digestion with Protease IV was done for 15 min at room temperature. Probes were hybridized for 2 h at 40 °C and washed twice in wash buffer (RNAscope® Wash Buffer Reagents, 310091). Amplification steps were performed by incubating with v2Amp1 (30 min), v2Amp2 (30 min) and v2Amp3 (15 min) at 40 °C with washes of 2 × 2 min in between steps. Sections were incubated with v2-HRP-C1 for 15 min at 40 °C and washed twice in wash buffer for 2 min. TSA-conjugated fluorophores were diluted 1:1,500 in TSA buffer (RNAscope® Multiplex TSA Buffer, 322809) and incubated for 30 min at 40 °C followed by 2 washes of 2 min and HRP blocker incubation for 30 min at 40 °C. The last steps were performed subsequently for v2-HRP-C2. Images were collected on Olympus FluoView FV1000 confocal microscope.

2.11. Data analyses

Data are expressed as mean ± SEM, except where otherwise noted. An equal number of male and female mice were used for each experiment and no sex differences were seen for any of the parameters measured. GraphPad Prism 8.0 (Graphpad software Inc. La Jolla, CA, USA) was used for statistical analyses and calculation of ED₅₀ values. All behavioral data were analyzed by one- or two-way ANOVA with repeated measures. Post-hoc comparisons were done with Bonferroni tests to correct for multiple comparisons. A P value of <0.05 was considered significant.

3. Results

3.1. MMG22 decreased mechanical hyperalgesia dose-dependently in neuropathic mice when delivered subcutaneously

Consistent with earlier reports [18,29,107], SNI produced robust mechanical hyperalgesia. Beginning at 5 days after nerve injury, mice displayed significantly increased paw withdrawal frequencies (compared to pre-surgical baseline measurements) which lasted the duration of the experiment. We evaluated the ability of MMG22, morphine, and loperamide to decrease mechanical hyperalgesia early (10 days after surgery / during the initiation phase) and late (30 days after surgery / during the maintenance phase) after nerve injury. MMG22 did not alter the percent withdrawal frequency of naïve animals or animals after sham procedure at any of the doses tested (data not shown). To increase the dynamic range of the assay, the filament chosen causes a 10% response frequency in naïve mice, because of this low baseline response, any analgesia caused by MMG22 may not have been detected in naïve or sham animals. Previous studies have shown little effect in naïve mice [2,3]. Subcutaneous administration of MMG22, morphine, and loperamide dose-dependently reduced mechanical hyperalgesia in nerve-injured mice (Fig. 1 A–C). Data are shown separately in mg/kg and nmoles/mouse for ease of comparison, but all data were analyzed together. We compared the ED_50s of MMG22, morphine, and loperamide early and late after nerve injury (two-way ANOVA; drug: [F_(2,183)= 4.8; P=0.0095], time: [F_(1,183) = 27.88; P<0.0001], time × drug: [F_(2,183)=12.54; P<0.0001], n=8–14). The potency of MMG22 was decreased late after nerve injury compared to early after nerve injury (P<0.0001), causing a rightward shift in the dose response curve (Fig. 1A, Table 1). Morphine did not exhibit the same shift in potency and was equipotent at reducing mechanical hyperalgesia early and late after nerve injury based on ED₅₀ values (Fig. 1B). MOR expression is known to decrease in the DRG following peripheral nerve injury [73,75,99,110,115,157,167], and this reduction has been linked to a concomitant loss of mu opioid responsiveness of DRG neurons [68]. To determine if loss of peripheral MOR expression after nerve injury was involved in the decrease in potency for MMG22, we repeated this experiment with the peripherally restricted MOR agonist loperamide. Loperamide does not penetrate the blood brain barrier at the doses tested [9,30,96,124], but has been shown to decrease nerve injury-induced hyperalgesia after systemic administration [25,45,126]. Consistent with previous data [45], there was no decrease in potency late after nerve injury as compared with the earlier time point (Fig. 1C). It has been shown that the bulk of MOR down-regulation in the DRG occurs within the first week after nerve injury [73,157], which suggests that any effects of MOR down-regulation on peripheral opioid potency would already be completed 10 days after nerve injury.

Figure 1. Cumulative dose response functions for reducing tactile hyperalgesia after spared nerve injury.

Mice were given subcutaneous injections of drug in increasing doses, early (10 days/open circles) or late (30 days/filled circles) after nerve injury. Paw withdrawal frequency to a 5.9 mN (0.6g) von Frey hair was measured 30 minutes after each injection and %MPE calculated based on pre-surgical baseline values. (A) MMG22 (blue) dose dependently reduced mechanical hyperalgesia in mice. The dose response curve is right shifted late after nerve injury compared to the earlier time point. (B,C) Morphine (red) and loperamide (purple) also dose dependently reduced mechanical hyperalgesia after nerve injury; however, no change in potency was observed over time. (D) After converting mg/kg into nmol/mouse, early (10 days) after nerve injury, MMG22 (blue) is more potent at reducing mechanical hyperalgesia than similar doses of morphine (red) or loperamide (purple). (E) Late (30 days) after nerve injury, the dose response curves for MMG22, morphine, and loperamide are overlapping. (F) MPEP (green) weakly reduced mechanical hyperalgesia early after nerve injury. There is a rightward shift in the dose response curve late after nerve injury. Data are presented on graphs as means ± SEM, n = 8–14 per group.

Table 1.

ED₅₀ values for MMG22, Morphine, Loperamide and MPEP early and late after nerve injury.

	nmol/mouse		mg/kg
Drug	Early	Late	Early	Late
MMG22	3.5 (1.4 – 6.6)a	219 (130 – 388)	0.12 (0.05 – 0.22)a	7.5 (4.5 – 13.3)
Morphine	139 (107 – 178)	232 (167 – 321)	1.6 (1.2– 2.0)	2.6 (1.9 – 3.7)
Loperamide	146 (122 – 174)	141 (114 – 173)	2.7 (2.3 – 3.1)	2.7 (2.2 – 3.3)
MPEP	5144 (3933 – 8210)b	28028 (16857 – 97792)	39.7 (30.4 – 63.4)b	216 (130 – 755)

All drugs given subcutaneously in a volume of 250ul.

^ap<0.0001 compared to morphine and loperamide (early and late) and MMG22 late (Two-way ANOVA followed by Bonferonni post hoc).

^bp<0.01 compared to MPEP late (Unpaired t-test). Data presented as mean (95% C.I.), n = 8–14 per group.

We converted the mg/kg doses into nmoles per mouse for direct comparison between drugs. Early after nerve injury, MMG22 was 40 times as potent as morphine or loperamide at reducing mechanical hyperalgesia (Fig. 1D, Table 1). Late after nerve injury however, all agonists were equipotent (Fig. 1E).

As one of the pharmacophores of MMG22 is MPEP, an mGluR₅ antagonist, we evaluated the ability of MPEP to decrease mechanical hyperalgesia early and late after nerve injury. At the highest dose, MPEP was unable to decrease mechanical hyperalgesia more than 50% early after nerve injury (Fig. 1F). This is consistent with previous reports showing that mGluR₅ antagonists alone have little effect [38,148,168] or no effect [53,56,153] on mechanical hyperalgesia after nerve injury. The ED₅₀ for MPEP was increased late after nerve injury compared to the earlier time point (unpaired t-test; t=2.73, P<0.01, n=8–12) (Fig. 1F, Table 1). The change in potency of MPEP over time after nerve injury suggests that this may contribute to the similar decrease in potency seen with MMG22.

3.2. Conditioned place preference in naïve and nerve-injured mice

To determine the potential rewarding properties of MMG22 we used the traditional CPP assay in naïve mice [10,145,146]. To examine the ability of MMG22 to promote reward by decreasing spontaneous ongoing pain, we used the variant aCPP assay in nerve-injured mice (both early and late after nerve injury) [28,48,65,137]. As a known drug of abuse, morphine was used as a positive control for naïve and nerve-injured mice. Loperamide and MPEP were used as positive controls for aCPP as they have both been shown to produce preference in nerve-injured animals but not in naïve or sham control animals [72,140]. Although SNI has been shown to decrease motor coordination in mice, it does not decrease overall motility [125] and therefore would not affect the development of preference for one chamber or another.

We tested the ability of 10 mg/kg s.c. MMG22 to induce CPP in naïve mice and aCPP in mice early and late after nerve injury. We compared the time mice spent in the drug paired chamber before and after conditioning (two-way ANOVA with repeated measures: injury: [F_(2,21)=5.39; P<0.05], conditioning: [F_(1,21)=5.77; P<0.05], injury X conditioning: [F_(2,21)=11.94; P<0.001], n=8–12). After 3 days of pairing with MMG22, naïve mice showed no preference for either chamber (P=0.73). The inability for MMG22 to induce place preference in naïve mice suggests that MMG22 may lack the addictive properties of traditional opioids. Early after nerve injury, mice spent more time in the drug paired chamber after conditioning with MMG22 (10 mg/kg) (P<0.0001). The ability for MMG22 to produce aCPP was limited to early after nerve injury as mice conditioned with MMG22 late after injury showed no increase in time spent in the drug paired chamber (P>0.99) (Fig. 2A). The ability of MMG22 to induce aCPP early after nerve injury, but not late, mirrors its decrease in analgesic potency over the same time period. A dose of 10 mg/kg MMG22 was also unable to induce place preference in sham operated animals early (10 days) after surgery (data not shown). Comparing the preference scores (one-way ANOVA; [F_(2,25)=12.41; P<0.001]) indicated that mice conditioned with 10 mg/kg MMG22 early after nerve injury had higher preference scores than naïve mice (P<0.001) as well as mice late after nerve injury (P=0.01) (Fig. 2B). There was no difference between the preference scores of naïve mice and mice tested late after nerve injury (P=0.24). The ability of MMG22 to induce aCPP only early after nerve injury suggests it is able to reduce spontaneous/ongoing pain early, but not late after nerve injury. The results parallel the decrease in potency of MMG22 for reducing mechanical hyperalgesia late after nerve injury.

Figure 2. Traditional and analgesic conditioned place preference in naïve and nerve-injured mice.

(A) Naïve mice show no difference in the time spent in the chamber paired with 10 mg/kg MMG22 before and after conditioning (pre: 837±66s vs post: 1183±89s; P=0.73). When paired with the same dose, nerve-injured mice spent significantly more time in the drug-paired chamber when conditioned early after nerve injury (pre: 788±40s vs post: 709±56s; P<0.0001) but not late after nerve injury (pre: 840±48s vs post: 1843±81s; P>0.99). (B) Mice had higher preference scores when paired with MMG22 early after nerve injury (322±56s) than naïve mice (−115±82; P<0.001), and mice late after nerve injury (49±52; P=0.01). Preference scores for naïve mice and mice conditioned late after nerve injury were not different (P=0.24). (C) Naïve mice spent more time in the drug paired chamber after conditioning with 10 mg/kg morphine (pre: 821±57s vs post: 1224±71s; P<0.0001). Nerve-injured mice also spent more time in the drug paired chamber after conditioning with 10 mg/kg morphine both early (pre: 802±58s vs post: 1147±54s; P<0.0001) and late (pre: 776±52s vs post: 1122±40s; P<0.0001) after nerve injury. (D) There is no difference between the preference scores of mice paired with morphine regardless of injury or timing (naïve: 403±62, early: 379±68, late: 346±48; P<0.99 for all comparisons). (E) Early after nerve injury, mice paired with 30 mg/kg MPEP mice spent more time in the drug paired chamber (pre: 795±71s vs post: 1045±40s, P<0.01). Naïve mice (pre: 854±40s vs post: 888±57s), and mice conditioned late after nerve injury (pre: 784±65s vs post: 799±68s) spent equal time in both chambers when conditioned with 30 mg/kg MPEP (P>0.99 for both). (F) Mice conditioned early after nerve injury (247±68) show higher preference scores for the drug paired chamber than naïve mice (−12±65; P<0.05) or mice conditioned late after nerve injury (5±39; P<0.05). There was no difference in preference scores between naïve mice and mice conditioned late after nerve injury (P>0.99). (G) Mice spent more time in the chamber paired with 10 mg/kg loperamide when conditioned early (pre: 757±31s vs post: 1025±76s; P<0.05) and late (pre: 774±43 vs post: 1001±51s; P<0.05) after nerve injury. Naïve mice showed no preference for the loperamide paired chamber (pre: 873±33s, vs post: 862±79s; P>0.99). (H) Mice conditioned early (268±84) and late (257±55) after nerve injury had larger preference scores for loperamide paired chamber than naïve mice (4±66; P<0.05 for both). Preference scores were not different for mice paired early or late after nerve injury (P>0.99). Data are presented as means ± SEM, n=8–12 per group. *P<0.05, **P<0.01, ***P<0.001, ****P<0.0001. One or two-way ANOVA followed by Bonferroni post hoc comparison.

Unlike MMG22, morphine at 10 mg/kg induced CPP in naïve mice (P<0.0001), as well as early (P<0.0001), and late (P<0.0001) after nerve injury (Fig. 2C) (two-way ANOVA with repeated measures; conditioning: [F_(1,21) = 0.11; P<0.0001], injury: [F_(2,21)=0.62; P=0.55], injury × conditioning: [F_(2,21)=0.32; P=0.73]). Comparison of the preference scores also showed no difference between morphine’s ability to produce place preference in naïve or nerve-injured mice (one-way ANOVA; [F_(2,22)=0.22; P=0.81]) (Fig. 2D). That morphine produced place preference equally well in naïve and nerve-injured mice (both early and late), suggests that at 10 mg/kg, the rewarding properties of morphine overshadow any increased reward the mice might get from the reduction of spontaneous ongoing pain.

We also compared the ability of MPEP and loperamide to induce CPP in naïve mice and aCPP in nerve-injured mice. The ability of MPEP to induce place preference differed between naïve and nerve-injured mice conditioned early and late after injury (two-way ANOVA with repeated measures; conditioning: [F_(1,24)=6.77; P<0.05], conditioning × injury [F_(2,24)=3.84; P<0.05], injury: [F_(2,24)=1.86; P=0.188]). Similar to previous studies [87,88], 30 mg/kg MPEP produced aCPP in mice conditioned early after nerve injury (P<0.01), but not CPP in naïve mice (P>0.99). Interestingly, 30mg/kg MPEP was also unable to produce aCPP in mice tested late after nerve injury (P>0.99) (Fig. 2E). Preference scores (one-way ANOVA; [F_(2,28)=5.9; P<0.01]) were higher for mice early after nerve injury than for naïve mice (P<0.05), or for mice late after nerve injury (P<0.05) (Fig. 2F). There was no difference between the preference scores of naïve mice and mice late after nerve injury (P>0.99). The temporary ability of MPEP to induce analgesic place preference after injury is consistent with the dose response data showing a significant decrease in potency of MPEP late after injury compared to the earlier time point.

The ability of loperamide (10 mg/kg) to induce place preference also differed among the groups (two-way ANOVA with repeated measures; conditioning: [F_(1,24)=10.47; P<0.01], injury: [F_(2,24)=0.13; P=0.88], conditioning × injury: [F_(2,24)=3.04; P=0.07]). Loperamide produced aCPP in mice tested early after nerve injury (P<0.05) as well as mice tested late after nerve injury (P<0.05). The same dose of loperamide was unable to produce CPP in naïve mice (P>0.99) (Fig. 2G). Comparing the preference scores (one-way ANOVA; [F_(2,29)=4.82; P<0.05]) revealed a difference between naïve mice and both groups of injured mice (P<0.05) but no difference between groups of mice after nerve injury (P>0.99) (Fig. 2H). The inability for loperamide to induce CPP in naïve mice is consistent with previous results [1]. Loperamide has also been shown to induce aCPP two weeks after nerve injury in rats [140]. That loperamide retained its ability to produce aCPP late after nerve injury also mirrors behavioral data showing no change in analgesic potency early vs late after nerve injury.

3.3. Dose dependency of analgesic place preference

We also examined the ability of different doses of MMG22 and morphine (1, 3 and 10 mg/kg) to induce analgesic place preference early after nerve injury. We first compared the time spent in the drug paired chamber before and after conditioning with MMG22 (two-way ANOVA with repeated measures; dose: [F_(2,21)=9.25; P=0.001], conditioning: [F_(1,21)=21.03; P<0.001], dose × conditioning: [F_(2,21)=8.29; P<0.01]). Early after nerve injury, mice paired with 3 mg/kg or 10 mg/kg MMG22 spent more time in the drug paired chamber (P=0.01, P<0.001 respectively). Mice paired with 1mg/kg MMG22 showed no preference for either chamber (P>0.99) (Fig. 3A). Similarly, preference scores also differed with dose (one-way ANOVA; [F_(2,23)=8.13; P<0.01], n=8–12). Preferences scores of mice given 3 mg/kg MMG22 and 10 mg/kg MMG22 - were greater than those produced by 1 mg/kg MMG22 (P<0.05 and P<0.01 respectively). There was no significant difference between preference scores of mice given 3 and 10mg/kg MMG22 (P>0.99) (Fig. 3B).

Figure 3. Analgesic place preference is dose dependent for MMG22, but not for morphine early after nerve injury.

(A) Early after nerve injury, mice spent more time in the drug paired chamber after conditioning with 3 mg/kg (pre: 895±31s vs post: 1173±75s; P=<0.01) or 10 mg/kg (pre: 837±66s vs post: 1182±89s,; P<0.001) or MMG22. Mice paired with 1 mg/kg MMG22 showed no preference for either chamber (pre: 757±72s vs post: 711±57s; P>0.99). (B) Mice paired with 3 mg/kg (278±82) and 10 mg/kg (322±56) MMG22 showed higher preference scores than mice paired with 1 mg/kg MMG22 (−46±73; P<0.05 and P<0.01 respectively). Preference scores for mice paired with 3 and 10 mg/kg MMG22 were not different from each other (P>0.99). (C) Early after nerve injury, mice spent more time in the drug paired chamber after conditioning with 1 mg/kg (pre: 829±7s vs post: 1106±6s; P=<0.01), 3 mg/kg (pre: 857±61s v post: 1149±58s; p<0.01), or 10 mg/kg morphine (pre: 802±58s vs post: 1147±54; p<0.001) (D) There were no differences in preference scores of mice paired with 1, 3, or 10 mg/kg of morphine (276±57, 292±88, 379±68 respectively; P>0.99 for all comparisons). Data presented as means ± SEM, n = 8–12 for all groups. *P<0.05, **P<0.01, ***P<0.001. One or two-way ANOVA followed by Bonferroni post hoc comparison.

Morphine at doses of 1, 3, and 10 mg/kg also produced place preference early after nerve injury (two-way ANOVA with repeated measures; dose: [F_(2,21)=0.15; P=0.87], conditioning: [F_(1,21)=53.7; P<0.0001], dose × conditioning: [F_(2,21)=0.25; P=0.79]). (Fig. 3C). Morphine increased the time mice spent on the drug paired chamber at 10 mg/kg (P<0.001) as well as at 3mg/kg (P<0.01). At 1mg/kg (~1/2 the ED₅₀) morphine was also equally effective (P=<0.01) at producing place preference in mice early after nerve injury. Examining the preference scores (one-way ANOVA: [F_(2,22)=0.61; P=0.55]) revealed no differences in preference scores between mice paired with any of the doses of morphine tested (P>0.99 for all comparisons) (Fig. 3D).

3.4. MMG22 did not alter locomotor activity

It is well known that MOR agonists induce hyper-locomotion [12,24,89,104,138] and thigmotaxis or “wall hugging” [6,51] in mice. However, there are conflicting data regarding the ability of mGluR₅ antagonists to alter locomotor behavior [81,85,168] and anxiolytic behavior [85,135,148] in rodents. We investigated whether s.c. MMG22 (10 mg/kg) altered either locomotor behavior or anxiety-like behavior in naïve mice and compared its effects to vehicle, morphine (10 mg/kg), MPEP (30 mg/kg) and loperamide (10 mg/kg). General locomotor activity indicated by the distance mice traveled during a 30 minute period in an activity chamber differed between groups (one-way ANOVA; [F_(4,35)=100.2; P<0.0001], Bonferroni post hoc pairwise against vehicle). Morphine caused an increase in distance traveled (P<0.0001) while loperamide had the opposite effect (P<0.05). Neither MMG22 nor MPEP had a significant effect on distance traveled (P>0.99, P=0.11 respectively) (Fig. 4A). The reverse pattern was observed for velocity (one-way ANOVA; [F_(4,40)=28.32; P<0.0001]) where morphine caused a decrease in velocity (P<0.0001) and loperamide increased velocity (P<0.05). Average velocities for mice after MMG22 and MPEP were not different from vehicle (P>0.99 for both) (Fig. 4B). We also examined the total time mice spent ambulatory during the same 30-minute sessions. Comparing treatments (one-way ANOVA; [F_(4,40)=116.3; P<0.0001]), morphine increased the total time mice spent ambulatory (P<0.0001) whereas loperamide had the opposite effect (P<0.05) (Fig. 4C). Total time spent ambulatory after MMG22 or MPEP were not different from vehicle (P>0.99, P=0.22 respectively).

Figure 4. Drug effects on locomotor activity and anxiety related behavior.

(A) Morphine increased the total distance traveled by mice over 30 minutes when compared to vehicle (P<0.0001). Loperamide had the opposite effect, decreasing the distance traveled by mice when compared to vehicle (P<0.001). Administration of MMG22 or MPEP had no effect on distance mice traveled compared to vehicle (P>0.99, P=0.11). (B) Morphine decreased the average velocity of mice compared to mice treated with vehicle (P<0.0001), whereas loperamide increased the average velocity slightly (P<0.05). The average velocity of mice treated with MMG22 and MPEP was not different than control mice (P>0.99 for both). (C) Morphine increased the amount of time mice spent ambulatory over a 30 minute period when compared to vehicle treated mice (P<0.0001). Treatment with loperamide had the opposite effect, showing a reduction in time mice spent ambulatory (P<0.05). MMG22 and MPEP had no effect of the time mice spent ambulatory (P>0.99, P=0.22 respectively). (D) Both morphine and loperamide decreased the time mice spent in the center zone of the chamber compared to vehicle treated animals (P<0.001 and P<0.05 respectively). MPEP increased the time mice spent in the center zone (P<0.0001). Treatment with MMG22 was not different from vehicle with respect to the amount of time mice spent in the center zone of the chamber (P=0.32). (E) When compared to vehicle, morphine caused a significant decrease in the number of entries mice made into the center zone (P<0.01). MPEP increased the number of times mice entered the center zone (P<0.001). Mice treated with MMG22 or loperamide entered the center zone the same number of times as vehicle treated mice (P=0.3, P=0.06 respectively). (F) Compared to vehicle, morphine decreased the distance traveled in the center zone (as a percentage of total distance traveled) (P<0.0001) while MPEP increased it (P<0.0001). Neither MMG22 nor loperamide had any effect on the distance mice traveled in the center zone (P>0.99 for both). (G) Representative traces of the locomotor behavior of mice over a 30 minute period after s.c. injection of vehicle, 10mg/kg MMG22, 10mg/kg morphine, 30mg/kg MPEP, or 10mg/kg Loperamide. Data presented as mean ± SEM, n = 9 per group. *P<0.05, **P<0.01, ***P<0.001, ****P<0.0001. One-way ANOVA followed by Bonferroni post hoc comparison.

We next examined the effect of each treatment on normal (non-pathological) anxiety-like behavior. Anxiolysis produced by drug treatment can be evaluated via the open-field test; anxiolytics will increase the amount of time rodents spend in the center region of an open field [114,129], as well as the number of entries made into the center region [114,128]. Because mGluR₅ antagonists have been shown to reduce anxiety like behavior in rodents [22,66,86,135,148], we determined if MMG22 produced anxiolytic activity. The amount of time mice spent in the central zone of the activity chamber differed among the treatment groups (one-way ANOVA; [F_(4,40)=27.30; P<0.0001]). Morphine reduced the time mice spent in the center of the chamber (P<0.001) as did loperamide (P<0.01). In agreement with previous studies, MPEP increased the amount of time mice spent in the center of the chamber (P<0.0001). Treatment with MMG22 did not differ from vehicle (P=0.32) (Fig. 4D). The number of times mice entered into the center zone of the chamber (one-way ANOVA; [F_(4,40)=24.25; P<0.0001]) was also decreased by morphine (P<0.001) and increased by MPEP (P=0.0001). Center zone entries after treatment with MMG22 or loperamide were not different from vehicle treated mice (P=0.3, P=0.06 respectively) (Fig. 4E).

Finally, to control for effects of hyper- or hypo-locomotion on the time spent in the center zone we compared the distance mice traveled in the center zone as a percent of total distance traveled after drug treatment (one-way ANOVA; [F_(4,40)=60.40, P<0.0001]). Compared to vehicle, morphine decreased the percentage of total distance traveled in the center of the chamber (P<0.0001). MPEP had the opposite effect and increased the percentage of total distance mice traveled in the center (P<0.0001). Treatment with MMG22 and loperamide were not different from control (P>0.99 for both) (Fig. 4F). The time mice spent ambulatory in the center zone (as a percent of total ambulatory time) followed the same pattern (data not shown).

Previous studies evaluating the effects of opioids on open field activity have shown that morphine does induce thigmotaxis “wall hugging” [6,51]; however the confounding hyper-locomotion induced by this dose of morphine makes interpreting this behavior difficult. The decrease in time spent in the central zone after loperamide was also unexpected. However, loperamide did not affect the distance mice traveled in the center area (as a percent of total distance traveled), suggesting that the decreased time spent in the center zone after loperamide may reflect the reduction in total time spent ambulatory and total distance traveled as opposed to anxiety related behavior. A representative trace of ambulatory data is shown in figure 4G. Over the course of the experiments it was also noted that MMG22, unlike morphine, did not induce Straub tail at any dose tested (data not shown). Cumulatively, the data suggests that MMG22 (at 10 mg/kg s.c.) does not affect locomotor behavior like traditional opioids, nor does it have the anxiolytic activity of mGluR₅ antagonists.

3.5. MMG22 did not cause respiratory depression

Whole body plethysmography was used to examine the effects of 3, 10 and 30 mg/kg of s.c. MMG22 and morphine on minute volume, respiratory frequency, and tidal volume in naïve mice (Fig. 5A–C, Table 2). Analysis of the AUC for minute volume normalized to baseline showed a significant effect of drug treatment (one-way ANOVA; [F_(6,35)= 4.77; P=0.001]). Post hoc comparison (pairwise against vehicle) showed a decrease in minute volume after systemic administration of 30 mg/kg morphine (P<0.05) (Fig. 5C). No other doses or treatments were different from vehicle. We also compared the effects of each drug on respiratory frequency [two-way ANOVA with repeated measures; drug: [F_(6,35)=10.53; P<0.0001], time: [F_(1,35)=29.03; P<0.0001], drug × time: [F_(6,35)=4.12; P=0.003]. Morphine decreased respiratory frequency at 3 mg/kg (P<0.05), 10 mg/kg (P<0.001) and at 30 mg/kg (P<0.001) (Table 2). Neither vehicle nor MMG22 had any effect on respiratory frequency (P>0.99 for vehicle and all doses of MMG22). Along with a decrease in respiratory frequency, morphine also caused a compensatory increase in tidal volume at 10mg/kg (P<0.001) and 30 mg/kg (P<0.001), but not at 3 mg/kg (P=0.24). MMG22 and vehicle had no effect on tidal volume (P>0.5) (two-way ANOVA with repeated measures; drug: [F_(6,35)= 1.77; P=0.15], time: [F_(1,35)=0.39; P=0.54], drug × time: [F_(6,35)=9.05; P<0.0001]) (Table 2).

Figure 5. MMG22 did not alter respiratory depression.

(A) Minute volume (frequency x tidal volume) before and after s.c administration of 3, 10 and 30 mg/kg morphine or MMG22. (B) Minute volume normalized to pre-drug baseline. Morphine dose dependently decreased minute volume in naïve mice. (C) AUC for minute volume is reduced after 30 mg/kg morphine (*p<0.05), but not MMG22. (n = 6 per group, data presented as mean ± SEM.) BL, baseline; AUC, area under the curve; s.c., subcutaneous. *P<0.05. One-way ANOVA followed by Bonferroni post hoc comparison.

Table 2.

Effects on respiratory frequency and tidal volume.

	Pre-drug baseline		20–30 Min post-injection
Drug / Dose	Frequency	Tidal Volume	Frequency	Tidal Volume	n
	(BPM)	(ml/breath)	(BPM)	(ml/breath)
Vehicle	318.05 ± 31.2	0.25 ± 0.02	294.91 ± 20.7	0.22 ± 0.02	6
MMG22 3 mg/kg	295.20 ± 31.1	0.28 ± 0.04	331.64 ± 19.4	0.25 ± 0.02	6
MMG22 10 mg/kg	350.19 ± 13.1	0.31 ± 0.02	356.16 ± 15.3	0.28 ± 0.02	6
MMG22 30 mg/kg	308.92 ± 32.8	0.23 ± 0.01	284.60 ± 14.6	0.23 ± 0.01	6
Morphine 3 mg/kg	345.13 ± 9.0	0.29 ± 0.03	326.69 ± 21.1*	0.30 ± 0.01	6
Morphine 10 mg/kg	288.51 ± 27.4	0.23 ± 0.01	200.92 ± 13.0***	0.28 ± 0.01***	6
Morphine 30 mg/kg	267.66 ± 16.8	0.22 ± 0.01	152.59 ± 4.3***	0.26 ± 0.01***	6

^*p<0.05,

^***p<0.001 compared to pre-drug baseline values (two-way ANOVA with repeated measures followed by Bonferroni post hoc). BPM, breaths per minute. Data presented as mean ± SEM, n = 6 per group.

3.6. Constipation produced by MMG22 and morphine

After 8 days of twice daily s.c. injection with either MMG22 (1 or 10 mg/kg), morphine (1 or 10 mg/kg) or vehicle we examined the effects of drug treatment on colonic motility. There was a significant difference between treatment groups in the time to bead expulsion (one-way ANOVA; [F_(4,25)=29.05; P<0.0001]). Both 1 and 10 mg/kg of MMG22 increased the time to bead expulsion compared to vehicle injection (P<0.05 and P<0.01 respectively) (Fig. 6). 10 mg/kg morphine greatly increased time to bead expulsion (P<0.0001 compared to all other treatments).

Figure 6. MMG22 and morphine decrease colonic motility.

MMG22 at 1 mg/kg and 10 mg/kg caused an increase in time to bead expulsion compared to vehicle (p<0.05, p<0.01 respectively). These effects were not different from 1 mg/kg of morphine. 10 mg/kg morphine decreased colonic motility significantly more than any other drug/dose combination (p<0.0001 vs vehicle, 1 mg/kg MMG22, 1 mg/kg morphine and 10 mg/kg MMG22). (n = 6 per group, data presented as mean ± SEM.) *P<0.05, **P<0.01, ***P<0.001, ****P<0.0001. One-way ANOVA followed by Bonferroni post hoc comparison.

3.7. mGluR₅ and MOR co-expression in mouse spinal cord and DRG

Previous studies have shown the MOR and mGluR₅ are expressed in primary afferent neurons [14,26,61], as well as in superficial dorsal horn neurons in the spinal cord [4,58,87,136,147,149]. To examine if both receptors are expressed together in the same neurons, we performed an RNAScope assay on lumbar spinal cord and DRG sections. Ten days after SNI surgery there was good expression of both MOR (OPMR1) and mGluR₅ (GRM5) mRNAs in the lumbar dorsal horn (Fig. 7A). Higher magnification images show that individual cells in the superficial (Fig. 7B) and deep (Fig. 7C) dorsal horn express both receptors. Individual cells in the DRG (Fig. 7D) also expressed the mRNAs for both receptors. High-resolution examination of the nuclear morphology (dapi stained) suggested the cells in question were neuronal (large nuclei, prominent nucleolar fading and folded nuclear membranes [41]; however, co-staining with cell type specific markers is needed for confirmation.

Figure 7. MOR (Opmr1) and mGluR₅ (Grm5) mRNAs co-expressed in dorsal horn and DRG 10 days after SNI.

(A-C) Examples of staining in the dorsal horn of SNI mice. Nuclei are stained using dapi (blue). (A) Low magnification image of the lumbar dorsal horn stained for MOR (OPRM1, right/red) and mGluR₅ (GRM5, middle/green), and overlay (right). Dashed line shows the grey-white matter boundary of the dorsal horn. (B) Higher magnification image of the superficial dorsal horn; OPMR1 (left/red), GRM5 (middle/green), overlay (right). Arrows indicate individual cells that co-express MOR and mGluR₅. (C) Higher magnification image of the deep dorsal horn; OPMR1 (left/red), GRM5 (middle/green), and overlay (right). Arrow indicates a cell that highly expresses both receptors. (D) High magnification image of a lumbar DRG; OPMR1 (left/red), GRM5 (middle/green) and overlay (right). Nuclei are stained using dapi (blue). Arrows indicate individual cells that express both receptors. Scale bar: 50 um for all images.

4. Discussion

Our results show that, in agreement with previous studies [107], MMG22 potently reduced mechanical hyperalgesia after nerve injury. For the first time, we show that MMG22 was able to reduce spontaneous pain, and importantly that systemic MMG22 lacked the rewarding properties and other centrally mediated side effects associated with traditional opioids. The bivalent design of MMG22 was intended to activate MOR and to inhibit mGluR₅ and to target a putative MOR-mGluR₅ heteromer. mGluR₅ has been characterized as a promising new target for chronic pain [94,95,152,153]. Activation of mGluR₅ produces pain [46,116,153] and mGluR₅ antagonists reduce pain behaviors in various pain models [32,37,61,86,148,152,168] without altering responses to acute noxious stimuli in naïve animals [123,152].

mGluR₅ is expressed on nociceptive primary afferents [53,64,95] and dorsal horn neurons [4,58,117,139,147,149]. Upregulation of mGluR₅ in the dorsal horn occurs in models of bone cancer pain [116], inflammatory pain [33,109], neuropathic pain [53,70,77,101], and peripheral polyneuropathies [141,158,165]. mGluR₅ upregulation was also reported in primary afferent fibers in models of neuropathic [53,67,76,158], and other injury-induced [34,154] pain.

Importantly, mGluR₅ modulates the analgesic and rewarding properties of opioids. mGluR₅ antagonists increase the potency of opioids [101,107,166], prevent analgesic tolerance [40,69,91,131,160,166], and decrease opioid reward [20,112]. The increased mGluR₅ expression in various pain conditions, combined with the analgesic profile of mGluR₅ antagonists, makes this receptor a promising target for pain management.

The exceptionally potent antinociception produced by MMG22 is due to MOR activation combined with antagonism of mGluR₅ and its co-receptor, NMDAR. Allosteric interaction of mGluR₅ via a covalent linkage with the NR2 subunit of the NMDAR has been shown to modulate neuronal excitability [15,19,106]. Pre-treatment of inflamed mice with an irreversible MOR antagonist (β-FNA) or the NMDAR antagonist (MK801) inhibited MMG22 induced analgesia [3], indicating a contribution of both pharmacophores of MMG22.

MMG22 was designed to target a putative MOR-mGluR₅ heteromer, which was supported by the relation between its linker length and potency [2,130]. Both MOR and mGluR₅ can form heteromers with other GPCRs [23,35,138] and MOR-mGluR₅ heteromers have been reported in vitro [122]. Bivalent ligands have increased affinity and selectivity for their targets [52,113], and antagonizing one receptor can enhance agonist-induced signaling at its heteromeric protomer [47]. Heteromer formation can be modulated by pathological states [43], and can alter signal transduction [50].

For the first time, we have shown that mRNAs for both receptors were co-expressed in neurons in the lumbar spinal cord and DRG early after nerve injury; suggesting potential targets for i.t. and systemic administration of MMG22 respectively. The co-expression of both receptors supports the potential for MOR-mGluR₅ heteromer formation in vivo. Previously, a non-overlapping pattern of MOR and mGluR₅ expression in the dorsal horn was found 8 weeks after nerve injury [107]. Once translated, the receptors may be trafficked to different cellular compartments. However, the pre- and post-synaptic location of the individual receptors would argue against this [8,13,58,139,149]. Alternatively, consistent with the 60-fold decrease in potency of MMG22 from early to late after injury (Table 1), the co-localization of its target receptors, and their heteromerization, may be transient and not present late after nerve injury. These possibilities may not be mutually exclusive, as studies have demonstrated that axonal targeting of mGluR₅ is dependent on the expression of Homer1a (an immediate early gene) [7], which is only transiently upregulated early after nerve injury [84]. This is the first study to demonstrate the colocalization of mRNAs for both target receptors in vivo.

4.1. Role of inflammation in the antinociception produced by MMG22

Early after nerve injury, MMG22 was 40 times more potent than morphine, whereas late after nerve injury the two were equipotent. Contrastingly, in a bone cancer pain model the analgesic potency of MMG22 increased in parallel with tumor growth and hyperalgesia [2,127,130]. The differences in potency of MMG22 may be explained by the timing and duration of the inflammatory response following nerve injury or tumor implantation. After nerve injury, there is an early pro-inflammatory response that is rapidly resolved after 2–3 weeks [5,83,143]. Whereas the inflammatory response after tumor implantation remains elevated over time [44,78,79,161]. The analgesic efficacy of MMG22 mirrors the time course of inflammation in bone cancer pain [127,130], and after nerve injury (in this study). A study of i.t. MMG22 after nerve injury did not report any statistically significant differences in potency overtime after nerve injury; however, the ED₅₀ was lowest at 7 and 17 days after injury and higher at earlier and later time points, also parallel to the time course of inflammation [107]. The combination of transient inflammation and persistent pain after nerve injury, allowed us to demonstrate that the potency of MMG22 induced analgesia is almost certainly dependent on ongoing inflammation, and not on ongoing pain.

A recent study showed a connection between the pro-inflammatory cytokine, tumor necrosis factor-α (TNF-α), and mGluR₅ upregulation after nerve injury [70]. TNF-α contributes to pain hypersensitivity [120,133,134,151]. A decrease in TNF-α levels reduced mGluR₅ expression and hyperalgesia, while intrathecal administration of TNF-α had the opposite effect [70]. The importance of TNFα for the development of neuropathic pain is consistent with previous studies showing a transient increase in TNF-α in the lumbar DRG [74,82,100,119] and spinal cord [74,159] after nerve injury. Upregulation of mGluR₅ may follow a similar timeline; with peak expression around 1–2 weeks post injury [67,98]. The potential involvement of inflammation in the potency of MMG22 is consistent with previous data showing that blocking astrocytes reduced the analgesic potency of intrathecal MMG22 [3].

Although our results show reduced potency for MMG22 late after nerve injury, many neuropathic pain conditions include more chronic inflammatory changes [111,132], suggesting that MMG22 may be a viable therapeutic for such conditions. For example, MMG22 provided long-term reduction in hyperalgesia produced by chemotherapy (manuscript in preparation), which has a persistent inflammatory component [164]. Future studies are needed to determine the types of neuropathic pain conditions that best respond to MMG22.

4.2. Addiction and abuse liability

The conditioned place preference assay, which assesses the rewarding attributes of a drug, is used as an initial screening tool for the addictive potential of drugs [11,145,146]. Morphine, but not MMG22, MPEP or loperamide, produced place preference in naïve animals. This strongly suggests that MMG22 lacks the abuse potential associated with traditional opioids. This is the first study to show that MMG22 is not rewarding in naïve mice.

MMG22 did produce analgesic place preference in animals conditioned early after nerve injury. The ability for an analgesic to produce place preference in injured animals reflects its ability to decrease ongoing, spontaneous pain [28,48,65,137]. The minimum dose of MMG22 required to produce aCPP was 3 mg/kg, a much higher dose than its ED₅₀ for reducing mechanical hyperalgesia (0.12 mg/kg), indicating that a higher dose of MMG22 is needed to decrease spontaneous pain as compared to evoked pain. MMG22 was also unable to produce aCPP late after nerve injury, when MMG22 showed a marked decrease in analgesic potency. The ability of MMG22 to produce place preference only early after nerve injury, when it is most potent at reducing hyperalgesia, supports MMG22 being an effective analgesic that lacks abuse potential. Additionaly, that MMG22 only induced aCPP when its analgesic potency was high, strongly suggests that the development of place preference for analgesics is related to the negative reinforcement of pain relief [65,92,137].

Unlike MMG22, morphine (at the doses tested) produced place preference in nerve-injured and naïve mice equally well. Studies in mice have shown that pain reduces the rewarding properties of opioids in ICR mice [90,102,103,150] whereas other studies have shown no changes or even enhancement of opioid reward in C57/BL6 mice [59,93,97]. The possibility that there are inherent strain differences in the processing of reward signals under conditions of pain warrants further exploration.

4.3. Side effect profile of MMG22

Systemic administration of MMG22 produced robust analgesia without the centrally mediated side effects of traditional opioids, including hyper-locomotion [12,104] and respiratory depression [16,105]. MMG22 had no effects on any of the locomotor behaviors measured or respiratory depression as suggested earlier [2,127]. A similar efficacy vs. side effect profile for the combination of the peripherally restricted MOR agonist loperamide and the delta opioid agonist oxymorphindole was reported [21], raising the possibility that MMG22 targets peripheral over central receptors. A simple reason for s.c. MMG22 targeting peripheral rather than spinal or supraspinal receptors is the relatively high molecular weight (852 daltons) of MMG22 [2]. Consistent with this possibility, MMG22 lacked the anxiolytic activity of the monovalent mGluR₅ antagonist MPEP, lacked the respiratory depressant effects which are known to be centrally mediated [39] and the main cause of opioid related overdose deaths [155], and was much less effective after intracerbroventricular administration [2]. Like traditional opioids, MMG22 caused constipation, a peripherally mediated side effect [156]. That MMG22 induced constipation, but none of the centrally mediated side effects associated with monovalent MOR agonists or mGluR₅ antagonists suggests that MMG22 may not pass the blood brain barrier when given systemically, and hence may be working peripherally to decrease mechanical hyperalgesia. Pharmacokinetic studies are needed to evaluate this possibility.